Hospitals Back Human‑Validated AI Over Autonomous Systems, New Survey Finds
The November Reaction Data poll released by Carta Healthcare highlights safety worries about black‑box tools.
Overview
- Nearly two‑thirds of respondents, 62.5%, identified misinterpretation of data as the top risk when AI runs without human oversight.
- Three‑quarters of organizations rely on human validation of AI outputs and rate clinician involvement in design and deployment as critically important.
- Only 12.5% said fully autonomous AI has delivered the most value to their organizations so far.
- Half of respondents stated AI should augment human decision‑making rather than replace clinical tasks.
- The findings come from a national online, opt‑in survey conducted in November 2025 by Reaction Data and released Dec. 16 by Carta Healthcare.